diff options
| author | saood06 <saood05@gmail.com> | 2025-05-09 02:09:59 -0500 |
|---|---|---|
| committer | GitHub <noreply@github.com> | 2025-05-09 10:09:59 +0300 |
| commit | bc6ae515ceb14eeaf198e00251a9689539cea176 (patch) | |
| tree | 82ca4aa8afa0dce38e0f3cb6fa9c7a78ec0065d2 /gguf-py/gguf/constants.py | |
| parent | 4084ca7331611da4426d781a15a6ffa68312759e (diff) | |
Support for Llama-3-Nemotron models (#377)
* conflict resolution
* Changes to make work and add longrope support
* Changes to n_attention_wv rule
* Untested support of 253B
* DeciLMCausalModel now reads rope_theta from config.json properly
* Remove errant Granite mentions
* Better n_attention_vw rule
* Update vocab.py
---------
Co-authored-by: Yee Man Chan <ymchan@gmail.com>
Co-authored-by: Iwan Kawrakow <iwan.kawrakow@gmail.com>
Diffstat (limited to 'gguf-py/gguf/constants.py')
| -rw-r--r-- | gguf-py/gguf/constants.py | 26 |
1 files changed, 26 insertions, 0 deletions
diff --git a/gguf-py/gguf/constants.py b/gguf-py/gguf/constants.py index d6feb9ea..e2f4eb1a 100644 --- a/gguf-py/gguf/constants.py +++ b/gguf-py/gguf/constants.py @@ -181,6 +181,7 @@ class GGUFType: class MODEL_ARCH(IntEnum): LLAMA = auto() + DECI = auto() FALCON = auto() BAICHUAN = auto() GROK = auto() @@ -316,6 +317,7 @@ class MODEL_TENSOR(IntEnum): MODEL_ARCH_NAMES: dict[MODEL_ARCH, str] = { MODEL_ARCH.LLAMA: "llama", + MODEL_ARCH.DECI: "deci", MODEL_ARCH.FALCON: "falcon", MODEL_ARCH.BAICHUAN: "baichuan", MODEL_ARCH.GROK: "grok", @@ -470,6 +472,26 @@ MODEL_TENSORS: dict[MODEL_ARCH, list[MODEL_TENSOR]] = { MODEL_TENSOR.FFN_DOWN_EXP, MODEL_TENSOR.FFN_UP_EXP, ], + MODEL_ARCH.DECI: [ + MODEL_TENSOR.TOKEN_EMBD, + MODEL_TENSOR.OUTPUT_NORM, + MODEL_TENSOR.OUTPUT, + MODEL_TENSOR.ROPE_FREQS, + MODEL_TENSOR.ATTN_NORM, + MODEL_TENSOR.ATTN_Q, + MODEL_TENSOR.ATTN_K, + MODEL_TENSOR.ATTN_V, + MODEL_TENSOR.ATTN_OUT, + MODEL_TENSOR.ATTN_ROT_EMBD, + MODEL_TENSOR.FFN_GATE_INP, + MODEL_TENSOR.FFN_NORM, + MODEL_TENSOR.FFN_GATE, + MODEL_TENSOR.FFN_DOWN, + MODEL_TENSOR.FFN_UP, + MODEL_TENSOR.FFN_GATE_EXP, + MODEL_TENSOR.FFN_DOWN_EXP, + MODEL_TENSOR.FFN_UP_EXP, + ], MODEL_ARCH.GROK: [ MODEL_TENSOR.TOKEN_EMBD, MODEL_TENSOR.OUTPUT_NORM, @@ -1149,6 +1171,10 @@ MODEL_TENSOR_SKIP: dict[MODEL_ARCH, list[MODEL_TENSOR]] = { MODEL_TENSOR.ROPE_FREQS, MODEL_TENSOR.ATTN_ROT_EMBD, ], + MODEL_ARCH.DECI: [ + MODEL_TENSOR.ROPE_FREQS, + MODEL_TENSOR.ATTN_ROT_EMBD, + ], MODEL_ARCH.BAICHUAN: [ MODEL_TENSOR.ROPE_FREQS, MODEL_TENSOR.ATTN_ROT_EMBD, |